Reverse engineering gene regulatory networks: coupling an optimization algorithm with a parameter identification...

Area of Science:

• Systems Biology
• Computational Biology
• Bioinformatics

Background:

• Inferring gene regulatory networks (GRNs) from time-series gene expression data is crucial for understanding biological systems.
• Key challenges include determining network structure and optimizing model parameters simultaneously.

Purpose of the Study:

• To develop an integrated approach for simultaneous network structure and parameter inference in GRNs.
• To address the limitations of existing methods that focus on only one aspect of network inference.

Main Methods:

• An iterative approach coupling parameter identification and optimization techniques.
• Parameter identification based on sensitivity measurements against internal perturbations.
• Hybrid Genetic Algorithm-Particle Swarm Optimization (GA-PSO) for parameter inference based on criticality.

Main Results:

• The approach was applied to diverse biological datasets (E. coli SOS repair, Rat CNS, yeast glycosylation).
• Successfully inferred solutions satisfying both network structure and behavior requirements.
• Demonstrated the ability to infer critical gene interactions and valid time expression profiles.

Conclusions:

• Network structure inference is challenging but critical for detailed biological networks.
• Parameter sensitivity analysis provides an indirect way to account for network structure.
• The integrated approach effectively infers gene interactions and expression profiles, advancing GRN analysis.